tion. The low conversions of 34 and 3-c to the other isomer and to product (as well as the respective quantum yields) ensure that the reactions observed are from the starting isomer only and that 4-c was not being destroyed as rapidly as it was formed in either case. The striking result is that the methyl migration of 3 is not stereospecific. This is in marked contrast to the divinylmethane process which proceeds stereospecifically with retention at C-l.3 Some basic difference between the two processes is indicated.'* Such nonspecificity could arise from (a) a stepwise process leading to long-lived biradical 5 which closes
Substituent Effects on T-T* Rearrangements. Methyl Migration in 0-tert-Butylstyrenes Sir: In a previous communication' we reported the effect of polar substituents on the photochemical rearrangement of 1,3-diarylpropenes to 1,2-diarylcy~lopropanes~ proceeding via aryl migration3 (eq 1). We noted' that
preferentially to 4-t; (b) reaction of 3-t and 3-c from a common (twisted) excited state; (c) a concerted .2, ,2, reaction from 3-c and a .2, ,2, reaction from 3-t.13 While no evidence is available to allow us to make a firm choice among these, a point against choice c is the result in the divinylmethane case where retention at C-1 is observed3with both cis and trans starting materials and where the steric requirements are quite similar to those in the present case. Nothing at present favors a over b though we feel more comfortable with a than b since we suspect that the twisting mentioned in the latter would be accompanied or closely followed by rapid deactivation to ground-state olefin. We note that while the divinylmethane rearrangement and the methyl migration show a basic difference in mechanism or timing of bonding, at least by this stereochemical criterion, the present results are similar to those found in the phenyl migration variation of the di-vmethane rearrangement, 4, l 5 where stereochemistry is apparently not retained and a common biradical intermediate seems likely. l 4 , I 6 A further similarity is seen in the effect of substituents on the two processes. l7
Acknowledgment is made to the donors of the Petroleum Research Fund, administered by the American Chemical Society, for support of this research. (12) Starting with either pure cis-1 or trans-1 we obtain only trans-2 at low conversion. However, due to the very low absorbance of cis-1 and its rapid isomerization to rrans-1 as well as the fact that the cyclopropanes elute from the gc on the tail of the cis-1 peak we cannot accurately determine a quantum yield for cis-1 -+ cyclopropane and eliminate the contribution from trans-1 to the total amount of cyclopropane formed. However, it appears here, too, the reaction is not stereospecific, and the results with 3 are not peculiar to the cyano substituent. (13) Combinations of a-c are also possible. (14) E. W. Valyocsik and P. Sigal, J. Org. Chem., 36, 66 (1971). (15) S. S. Hixson, J . Amer. Chem. Soc., 94, 2507 (1972). (16) However, it should be noted that time dependence studies, which show that the lack of stereospecificity was truly a result of kinetic control, were not presented in ref 14. (17) S. S. Hixson and 7. P. Cutler, J . Amer. Chem. Soc., 95, 3032 ( 1973). (18) Undergraduate honors research participant.
Stephen S. Hixson,* Timothy P. Cutler1* Department of Chemistry, Utzicersity of Massachusetts Amherst, Massachusetts 01002 Received December 14, 1972
Journal of the American Chemical Society / 95:9 / M a y 2, I973
when the migration terminus is substituted with a strongly electron withdrawing group (X = CN; Y = H), the reaction proceeds rapidly (krel = 41) relative to the unsubstituted compound (X = Y = H). However, substitution by an electron donating group (X = OCH3; Y = H) drastically curtails the rate (h6 0.04). Such a polar effect was striking and unexpected; it indicated that the electron density at the migration terminus was a very important factor in determining the rate of phenyl migration, much as in the ground-state migration of phenyl t o an adjacent carbon. Little is known about the nature of the electronic interactions that bring about rearrangements in A-T* systems and, in particular, the effect of a-electron distribution on these rearrangements. Thus, we have extended the studies noted above on the migration of unsaturated carbon in an excited styryl system to include a study of substituent effects on the migration of saturated carbon occurring upon photolysis of @-tertbutylstyrene4 (eq 2). Such a study promised to pro-
3a, X = H
b, X = C N c,
b, X = C N C,
vide a further5 comparison of the two types of rearrangement, one which can proceed via bridging between two a systems, the other which cannot. In addition, though, it would provide an indication of whether the effects noted above were general and applied to different sorts of migration processes. The irradiation6 of 3a was followed closely by gas chromatography. As reported by Griffin in his original study of 3a,4 small amounts of 4a (cis and trans) were initially formed but did not accumulate due to their rapid further transformation. In the present case the (1) S. S. Hixson, J . Amer. Chem. SOC.,94, 2507 (1972). (2) G. W. Griffith, J. Covell, R. C. Petterson, R. M. Dodson, and G. Klose, ibid., 87, 1410 (1965). (3) S . S. Hixson, Tetrahedron Lett., 1155 (1972). (4) H. Kristinsson and G. W. Griffin, J . Amer. Chem. Soc., 88, 378 (1966). (5) S. S. Hixson and T. P. Cutler, J . Amer. Chem. Soc., 95, 3031 ( 1973). (6) Synthetic and experimental details will be reported in a full paper.
photolysis was stopped when the amount of 4a appeared to be at a maximum. Olefins were then removed by permanganate oxidation; 4a (cis and trans) was isolated by column chromatography and identified by comparison (nmr and ir spectra; gc retention times) with an authentic sample of 4a (cis and trans). We point out that this is the first isolation of 4a from 3a although the work of Griffin4 left little doubt as to its being formed. Similarly, cyclopropanes 4b (cis and trans) could be isolated from the photolysis of 3b.j Strikingly, however, irradiation of 3c led to no observable formation of 4c or of any other product. Indeed, prolonged irradiation gave solely trans-cis isomerization with only a very slow disappearance of starting material being noted. Triplet sensitization of 3a-c with xanthone gave only cis-trans isomerization indicating the observed migrations are singlet state processes. Quantum yield studies of the rearrangements of the trans olefins were carried out on a merry-go-round apparatus using 254-nm light. At the low conversions ( of these irradiations only trans cyclopropanes were f ~ r m e d . The ~ results are shown in Table I. To Table I. Quantum Yield and Relative Rate Data for Olefins la-c
3a 3b 3c